Title :
Tanh-like Activation Function Implementation for High-performance Digital Neural Systems
Author :
Marra, S. ; Iachino, Maria A. ; Morabito, Francesco C.
Author_Institution :
Dept. of Informatics, Math., Electron. & Transp., Reggio Calabria Univ. "Mediterranea"
Abstract :
In this paper, a high-performance implementation of a programmable tanh-like activation function is presented. The proposed function can be successfully used to train neural networks obtaining generalization ability slightly better than standard activation functions. The achieved accuracy, the high computational speed and the small amount of area resources make the proposed solution ideal for digital implementations
Keywords :
digital arithmetic; logic design; neural chips; programmable circuits; digital neural systems; generalization ability; neural networks; programmable tanh-like activation function; Arithmetic; Artificial neural networks; Computer networks; Energy consumption; Field programmable gate arrays; Hardware; Informatics; Integrated circuit interconnections; Neural networks; Neurons;
Conference_Titel :
Research in Microelectronics and Electronics 2006, Ph. D.
Conference_Location :
Otranto
Print_ISBN :
1-4244-0157-7
DOI :
10.1109/RME.2006.1689940
